A

PROJECT PHASE-I REPORT

ON

Automated Financial Document Processing System using ML

SUBMITTED IN THE PARTIAL FULFILLMENT OF THE REQUIREMENTS

FOR THE AWARD OF THE DEGREE

OF

BACHELOR OF TECHNOLOGY (COMPUTER ENGINEERING)

SUBMITTED BY

Gaurav Deshmukh BCOA35

Kunal Salunkhe BCOA46
Pranjali Bansode BCOA59
Pravin Patrike BCOA64

Under the Guidance of

Mrs. Sunita Nandgave

DEPARTMENT OF COMPUTER ENGINEERING

G. H. RAISONI COLLEGE OF ENGINEERING AND MANAGEMENT

(An Autonomous Institute affiliated to SPPU)
WAGHOLI, PUNE-412207

1
 2022 -2023

CERTIFICATE

This is to certify that the project report entitles

“Automated Financial Document Processing System using ML”

Submitted by

Gaurav Deshmukh BCOA35

Kunal Salunkhe BCOA46
Pranjali Bansode BCOA59
Pravin Patrike BCOA64

are bonafide students of this institute and the work has been carried out by them under the
supervision of Mrs. Sunita Nandgave and it is approved for the partial fulfillment of the
requirement of Savitribai Phule Pune University, for the award of the degree of Bachelor of
Technology (Computer Engineering).

Mrs. Sunita Nandgave Ms. Gayatri Bedre

Guide Project Coordinator

Dr. Simran Khiani Dr. R.D. Kharadkar

HOD Director

Place: Pune
Date:

ACKNOWLEDGEMENT
2
 We deem it a pleasure to acknowledger our sense of gratitude to our project guide Mrs.
Sunita Nandgave under whom we have carried out the project work. His incisive and objective
guidance and timely advice encouraged us with constant flow of energy to continue the work We
wish to reciprocate in full measure the kindness shown by Dr. Simran Khiani (H.O.D, Computer
Engineering) who inspired us with her valuable. suggestions in successfully completing the print
work
We shall main grateful to all teaching staff and non-teaching staff, for providing us a strong
academic atmosphere by enforcing strict discipline to do the project work with utmost
concentration and dedication.
Finally, we must say that ne height is ever achieved without sunie sacrifices made at someend
and it is here where we owe our special debt to our parents and our friends for showing their
generous love and care throughout the entire period of time.

NAME OF THE STUDENTS & SIGNS

Signature :
Name : KM
Radha
Roll No :
1809510037
Signature :
Name : KM
Radha
3
Roll No :
1809510037
1. Signature:
Name: Gaurav Deshmukh
Roll No: BCOA35

2. Signature:
Name: Kunal Salunkhe
Roll No: BCOA46

3. Signature:
Name: Pranjali Bansode
Roll No: BCOA59

4. Signature:
Name: Pravin Patrike
Roll No: BCOA64

ABSTRACT
In today’s world technology has affected each sector. But still, within the banking sector, we
tend to face several problems. Bank handles giant volumes of cheques within the clearing
method. The method involves several technical verifications as well as signature verification. A
number of these steps are manual and need human intervention to finish the technique. This
method needs high human capital preparation and longer interval. Fallacious practices are one in
every one of the foremost problems plaguing banks in recent times. Hiring a lot of folks to
manage the cheque verification method does not solve this issue. This project helps as an answer
to the present downside. Optical Character Recognition (OCR), workflow systems, and machine
learning techniques are the key technologies to building automatic document processing. When it
involves the clearance of bank cheques and financial transactions, this alternate methodology for
the process of bank cheques with stripped-down human intervention additionally saves time, and
automating the method through computer vision technology can facilitate ensuring that solely
authentic banknotes are handled and under such scenario, cheques will not be delayed in
reaching its destination. We tend to propose an automated system that extracts relevant details of

4
a bank cheque like Payee Name, Amount, Date, and Bank Name using Optical Character
Recognition and Deep Learning and verifies the signature on the cheque with the prevailing
signature keep within the database using feature extraction and principal element analysis. Our
innovation aims to learn the banking industry by re-innovating the other competent chequebased
financial transaction system that needs machine-driven system intervention.

TABLE OF CONTENTS

CHAPTER TITLE PAGE NO.

[Link]. Title Of Chapter Page No.

01 Introduction 9
1.1 Problem Definition 9
02 Literature Survey 10
03 Hardware & Software Requirements 12
04 Feasibility Study and System Design 13
4.1 Feasibility Study 13
4.1.1 Technical Feasibility 13

5
4.1.2 Operational Feasibility 13
4.1.3 Economic Feasibility 13
4.2 System Design 14
4.2.1 Activity Diagram 14
4.2.3 Sequence Diagram 14
4.2.4 Flow Chart 15
05 Module Implementation and System Integration 16
5.1 Data Collection 16
5.2 Data Preprocessing 16
5.3 Model Implementation 16
5.3.1 Methodology 16
5.3.2 Implementation 17
06 Other Specification 19
6.1 Advantages 19
6.2 Limitation 19
6.3 Applications 19
07 Conclusions 21
08 References 22

LIST OF ABBREVATIONS

ABBREVIATION ILLUSTRATION

ML Machine Learning
CNN Convolutional Neural Network
OCR Optical Character Recognition

6
 LIST OF FIGURES
FIGURE ILLUSTRATION PAGE NO.

4.2.1 Activity Diagram 14

4.2.2 Sequence Diagram 15
4.2.3 Flow Chart 15
5.3.2 GUI Diagram 17

7
 LIST OF TABLES
TABLE ILLUSTRATION PAGE NO.

2.1 Literature Survey 10

3.1.1 Hardware Requirements 12
23.1.2 Software Requirements 12

8
 CHAPTER I
1.1 INTRODUCTION
In the modern world, technology has an impact on every industry. Still, there are numerous
problems in the financial industry. Large numbers of checks are handled by the bank through
the clearing process. The process includes signature verification along with a number of
technical checks. A few of these processes are manual and require human assistance to be
completed. The current approach requires extensive human capital preparation and a longer
time frame. One of the main issues affecting banks recently is false practises. This problem
cannot be resolved by hiring more personnel to oversee the check verification procedure.
This initiative contributes to the resolution of this issue.

9
 The main technologies for developing automatic document processing include optical
character recognition (OCR), workflow systems, and machine learning strategies. This
alternative method for processing bank checks with minimal human involvement saves time
when it comes to clearing financial transactions and bank cheques. Automating the process
through computer vision technology can help ensure that only genuine banknotes are
handled, and in that case, checks won't be delayed in reaching their destination
We tend to recommend an automated system that uses feature extraction and principal
element analysis to compare the signature on the check to the current signature stored in the
database and uses optical character recognition and deep learning to extract crucial
information from a bank check, such as the Payee Name, Amount, Date, and Bank Name. By
redesigning another competent check-based financial transaction system that requires
machine-driven system intervention, our innovation intends to get insight into the banking
industry.

1.2 PROBLEM DEFINITION

The automated financial documents processing system using ML can help reduce the time
and effort required to process financial documents. The system can be trained to
automatically read and categorize financial documents, extract key information, and generate
reports.

CHAPTER II
2.1 LITERATURE SURVEY

[Link] Paper Title Authors Name Year Problem Discussed/

Contribution

10
1 Cheque Processing Using Ashok T. Gaikwad 2019
OCR Ali Mansour
Aims to provide
Precise scanning of
financial documents.

2 Image Processing Based T Vairam 2016

Document Scanner. S Sarathambekai
R Balaji
The objective of the paper
is to create a document
scanner which
automatically detects
details using technology

3 Cheque Processing Wenbin Zhang 2016

System Yuan Yuan Aims to provide a precise
Yanyan Hu cheque scanner which
automatically detects and
verfies details.

4 Handwritten Cheque Rajib Ghosh 2018

Recognition in Bank Chinmaya Panda Aims to develop a bank
Cheques Prabhat Kumar cheque handwritten text
recognition system for
recognizing the
handwritten characters.

5 Bank Cheque J. Girish 2015

Identification P. Geetha
Aims to provide image
recognition and object
detection.

11
6 Bank Cheque Signature Harshit Sharma 2019
Verification Using SIFT
To verify bank cheque
signature by using scale
invariant feature transform
(SIFT) algorithm.

CHAPTER III

3.1 HARDWARE AND SOFTWARE REQUIRMENTS

3.1.1 HARDWARE REQUIREMENTS:

12
3.1.2 SOFTWARE REQIREMENTS:

CHAPTER IV
FEASIBILITY STUDY AND SYSTEM DESIGN
4.1 FEASIBILITY STUDY
The feasibility of the project is analyzed in this phase and business proposal is put forth with
a very general plan for the project and some cost estimates. During system analysis the
feasibility study of the proposed system is to be carried out. This is to ensure that the
proposed system is not a burden to the college/organization. For feasibility analysis, some
understanding of the major requirements for the system is essential.
Three key considerations involved in the feasibility analysis are
• Technical feasibility
• Operational feasibility
• Economic feasibility

13
4.1.1 Technical Feasibility
Technical feasibility evaluates the technical complexity of the document processing
System and often involves determining whether the System can be implemented with
state-of-the art techniques and tools. In the case of Financial document processing, an
important aspect of technical feasibility is determining the model to which the GUI
application will be integrated. The model used for computation of images is an important
determinant to software's quality and makes it vital to the application success.

4.1.2 Operational Feasibility

Operational feasibility refers to the ability of a system to perform all its operations
effectively and efficiently. Our project provides an easy-to-use user-friendly GUI, so the
user can perform all his/her task with few clicks.

4.1.3 Economic Feasibility

Our project is economically feasible because we consider to buy bare minimum hardware
for working of this project.

4.2 SYSTEM DESIGN

4.2.1 ACTIVITY DIAGRAM
The system receives the check, and as part of the pre-processing process, OCR is
initially performed on the entire set of data. Currently, we have created templates for
banks where customers may input checks so that data can be retrieved using OpenCV to
rapidly and effectively crop out the necessary portions. After receiving the results, we use
string matching and manipulation techniques to extract the specific bank template from
the result. The process of establishing a system’s architecture, parts, modules, interfaces,
and data in order to meet predetermined criteria is known as system design. You may
think of systems design as the application of systems theory to the creation of products.
The most popular methodologies for designing computer systems are increasingly those
that use object-oriented analyses and procedures

14
 Figure 1 Activity Diagram

4.2.2 SEQUENCE DIAGRAM

Figure 2 Sequence Diagram

15
 4.2.3 Flow chart

the cheque is fed to the system which performs OCR at first stage on whole data as a pre
processing step. data can be extracted using OpenCV to crop out the required parts effectively
and efficiently. After getting the results we perform string matching and manipulation methods
on the result extracted to get the specific bank template. System design is the process of the
defining the architecture, components, modules, interfaces, and data for a system to satisfy
specified requirements. Systems design could be seen as the application of systems theory to
product development. Object-oriented analysis and methods are becoming the most widely used
methods for computer systems design. Systems design is therefore the process of defining and
developing systems to satisfy specified requirements of the user.

CHAPTER V
MODULE IMPLEMENTATION AND SYSTEM INTEGRATION
5.1 DATA COLLECTION
Data collection is the process of gathering and measuring information on targeted variables
in an established system, which then enables one to answer relevant questions and evaluate
outcomes. Data collection is a research component in all study fields, including physical and
social sciences, humanities, and business. While methods vary by discipline, the emphasis on
ensuring accurate and honest collection remains the same. The goal for all data
collection is to capture quality evidence that allows analysis to lead to the formulation of
convincing and credible answers to the questions that have been posed. We have collected
data from various resources for best possible training of our model,

1.2 DATA PRE-PROCESSING

16
 Some pre-processing is applied on the image to reduce noise. There are some common
methods of preprocessing: Smooth, Dilate, Erode, Median, Open, Close etc. To maximize the
efficiency of automated document processing, document images first need to be improved
through things such as noise reduction, de-skewing to straighten the documents, and
temporarily eliminating non-text elements like logos, pictures, etc. Computer vision can be
an element of this pre-processing step, as it helps computers perform actions that are
extremely easy for humans, but not so easy for a machine.
In the realm of automated document processing, computer vision (CV) helps process
document images for very accurate optical character recognition (OCR).

1.3 MODEL IMPLEMENTATION

1.3.1 METHODOLOGY
From The bank check verification tool demanded the completion of a few crucial steps in a
precise order. The tool first checked the validity of the IFSC on the check, then it checked the
check number to see if the check came from the set of check booklets sent to the account
holder or not. The amount was next confirmed by ensuring that the customer had a sufficient
amount in his account to cover the release of the check, and finally the validity of the cheque
issuer(s)' signatures. The clearance of the check for the transfer or withdrawal of funds was
only granted after the conclusion of these crucial verification stages.

1.3.2 IMPLEMENTATION
 Computer vision: - OCR is used for image processing and machine learning
software library which is used in this project for image processing purpose and
understanding the data And applied CNN to recognize characters and numbers.

 Graphical User Interface (GUI): -The GUI contains all the options required for the
maintenance of the software. This can be used for management and other
troubleshooting objectives.
- the GUI is built by HTML and CSS which makes up a beautiful and easy to
understand dashboard from where doing necessary action for user becomes
convenient

17
Figure 5 GUI Diagram

18
 CHAPTER VI
OTHER SPECIFICATION
6.1 ADVANTAGES
1. Speed: An automated system can process financial documents much faster than a human
can. This can free up time for your employees to focus on other tasks.

2. Accuracy: Automated systems can reduce the chances of errors in financial document
processing. This can save your business money and help to avoid potential penalties
from regulators

3. Flexibility: An automated system can be customized to meet the specific needs of your
business. This flexibility can make it easier to scale your business in the future..

6.2 LIMITATIONS
While machine learning can be extremely effective for automating the processing of
financial documents, there are some limitations to keep in mind. First, machine learning
algorithms require a large amount of data to train on in order to be effective. If you don't
have enough data, the algorithm will not be able to learn and accurately process your
documents. Second, machine learning algorithms can be expensive to train and run. If you're
on a tight budget, you may not be able to afford the upfront cost of training the algorithm.
Finally, machine learning is not perfect and there will still be some errors made when
processing documents. However, these errors can usually be corrected with human
intervention.

19
6.3 APPLICATIONS

 Greater data security: storing data online is more secure than physical or local
storage. Electronic documents also make audit trails easier to perform and keep your
data safe from fire, flood, and theft.
 Higher data accuracy: data is extracted with minimal human intervention to prevent
potential mistakes.
 Better GDPR compliance: documents are easily indexed and retrieved within the 30-
day time window of data requests.
 Improved data control: document lifecycles can be monitored closely.
 Improved workflow: documents are processed, searched for and retrieved quickly,
giving staff more time to focus on core business tasks.
 Reduced cost: money can be invested elsewhere instead of on document processing
labor, paper cost, and data mistakes.
 Better sustainability compliance: for those working towards more sustainable offices,
document automation means less paper and less physical filing space needed.
 Faster turnaround times: payments can be made faster and with fewer mistakes

20
 CHAPTER VII
CONCLUSIONS
7.1 CONCLUSIONS
Using OCR, CNN, SIFT, and SVM, we created a model to validate bank checks.
We conducted CNN to provide accurate output for the handwritten figures printed on the
cheque book let, while we utilised OCR to accurately and efficiently detect the machine
typographic characters. For the purpose of verifying the check, which is one of the main
causes of its bounce off and the halting of financial transactions, we devised and
implemented the technique to transform numbers into words. This was accomplished
using an OCR approach to identify the machine-printed digits and 97.7 percent exact
[Link] network was then trained using a separate database, and after training it to
the appropriate degree of accuracy, we tested the trained model by comparing numbers to
words using multiple data sets. Evidently, we improved upon the prior accuracy of 99.05
percent for the CNN employed for digit recognition by achieving an accuracy of 99.14
percent for digit recognition. In a similar vein, we were successful in achieving an
accuracy of up to 99.94 percent for character recognition using CNN. We employed SIFT
and SVM classifiers to identify the relevance for signature verification and obtained 98.1
percent accuracy in the process of recognising signatures. We employed SIFT and SVM
classifiers to identify the relevance for signature verification and obtained 98.1 percent
accuracy in the process of recognising signatures. By reducing time and labour compared

21
 to the traditional and manual technique, our invention will assist financial institutions in
automating the process of cheque clearance. In order to automate the process of verifying
bank checks based on the recognition of the handwritten numeric digits, we completed
the verification of bank checks using CNN in our study work. But in order to effectively
clear bank checks, we also need to validate bank checks and be able to work online with a
database that can answer the necessary queries in real time

CHAPTER VIII
REFERENCES
8.1 REFERENCES
[1] Sook Chin Chiew , Xin Yuan Law , Ren Zhang Tan , XinYing Chew, Khai Wah Khaw
“Digital Recognition by Deep Learning Techniques: A Proposed Digit Recognizer to Automate
Cheque Deposition”,In Amity Journal of Computational Sciences (AJCS) 2019

[2] Ganga Holi and Divya K. Jain, “Convolutional Neural Network Approach for Extraction and
Recognition of Digits from Bank Cheque Images”, In International Conference on Intelligent
Sustainable Systems, 2019

[3] Mukesh Jha, Madhur Kabra, Sahil Jobanputra, and Prof. Rupali Sawant, “Automation of
Cheque Transaction using Deep Learning and Optical Character Recognition”, In Second
International Conference on Smart Systems and Inventive Technology (ICSSIT 2019)

[4] Saleem Ulla Shariff, Maheboob Hussain, Mohammed Farhaan Shariff, “Automated bank
cheque verification using image processing and deep learning methods”, Springer
Science+Business Media, LLC, part of Springer Nature 2020

[5] Victor Carbune, Pedro Gonnet, Thomas Deselaers, Henry A. Rowley, Alexander Daryin,
Marcos Calvo, LiLun Wang, Daniel Keysers, Sandro Feuz, Philippe Gervais, “Fast multi-
language LSTM-based online handwriting recognition”, International Journal on Document
Analysis and Recognition (IJDAR) (2020)

[6] Girish C. J, Mrs. Geetha G. P “Design of Bank Cheque Validation System”, International
Journal of Engineering Research & Technology (IJERT) 05, May-2015

22
[7] Mohit Mehta, Member, IACSIT, Rupesh Sanchati and Ajay Marchya, “Automatic Cheque
Processing System”, International Journal of Computer and Electrical Engineering, 2018

[8] Sebastian Salazar-Colores, Eduardo Cabal-Yepez, “A Fast Image Dehazing Algorithm Using
Morphological Reconstruction” IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL. 28,
NO. 5, MAY 2019 [9] Beresneva A, Epishkina A, Shingalova D (2018) Handwritten signature
attributes for its verification. In: IEEE conference of Russian young researchers in
electrical and electronic engineering (EIConRus), pp 1477–1480

[9]Uhliarik, I.: Handwritten character recognition using machine learning methods. Bachelor’s
Thesis, Department of Applied Informatics, Bratislava (2013)

[10]Vision Tech: OCR neural networks and other machine learning techniques
Google Scholar

[11]Fanany, M.I.: Handwriting recognition on form document using convolutional neural

network and support vector machines (CNN-SVM). In: 2017 5th International Conference on
Information and Communication Technology (ICoIC7), Melaka, pp. 1–6 (2017).

[12]Shih, Y., Wei, D.: Handwritten Sanskrit recognition using a multi-class SVM with K-NN
guidance. MIT

[13]LeCun, Y., Boser, B., Denker, J.S., et al.: Backpropagation applied to handwritten zip code
recognition. Neural Comput. 1, 541–551 (1989

[14]Coates, A., Carpenter, B., Case, C., et al.: Text detection and character recognition in scene
images with unsupervised feature learning. In: ICDAR 2011, Computer Science Department,
Stanford University, Stanford, CA, USA. Nov 2011

[15]Alam Miah, M.B., Yousuf, M.A., et al.: Handwritten courtesy amount and signature
recognition on bank cheque using neural network. Int. J. Comput. Appl.(0975–0887) 118(5)
(2015)

[16]Suresh Kumar Raju, C., Mayuri, K.: A new approach to handwritten character recognition.
Int. J. Innov. Res. Comput. Commun. Eng. 5(4), 9003–9012 (2017

[17]Mehta, M., Sanchat, R., et al.: Automatic cheque processing system. Int. J. Comput. Electr.
Eng. 2(4) (2010).

[18]Narkhede, S.G., Patil, D.D.: Signature verification for automated cheque authentication
system based on shape contexts. Int. J. Comput. Sci. Inform. Technol. 5(3), 3297–3300 (2014)

23